UC Berkeley Phonology Lab Annual Report (2008) Variation in Palatal Production in Buenos Aires Spanish Charles B. Chang* University of California, Berkeley 1. Introduction Argentine Spanish is well-known among varieties of Spanish for its “hardened” palatals. The sonorant palatals of European Spanish have undergone fortition in most dialects of Argentine Spanish, resulting in fricatives or affricates in all environments. In one recent study focusing on Buenos Aires, Colantoni (2006) examined the change from sonorant /j/ to obstruent /ʒ/ in some detail. However, when walking through Buenos Aires today, one actually hears voiceless [ʃ] more often than voiced [ʒ], a fact that might be behind marked discrepancies among layperson impressions of these sounds. On the one hand, many popular authors indicate that the predominant pronunciation of palatals in Argentina is voiced. For instance, Bao and Greensfelder (2002: 187) claim that “most Argentines say ‘zh’ for [‘y’] and for ‘ll’”, while others make reference to the “replacement of the ‘ll’ and ‘y’ sounds by a soft ‘j’ sound, as in ‘beige’” (Dilks 2004: 25) and to “the unique way in which porteños [Argentines from Buenos Aires] pronounce the letters ll and y (with a sound akin to the French ‘je’)” (Tozer 2001). Very recent publications present similar thoughts, stating that “Argentine Spanish differs from other countries in its pronunciation of the ll and y not as a y but as English speakers would pronounce a j” (Luongo 2007: 59). Lawrence (2007: 367) concurs that “the local accent imposes a soft ‘j’ sound on the ‘ll’ and ‘y’”, such that “llave (key) sounds like ‘zha-ve’ and desayuno (breakfast) sounds like ‘de-sa- zhu-no’” (Luongo et al. 2007: 421). On the other hand, some authors suggest that the predominant pronunciation of palatals in Argentina has become either partially or completely voiceless. Greenberg (2000: 15) asserts that “the ‘y’ becomes a combination of ‘z’ and ‘sh’”, a statement echoed by McCloskey (2006: 395–396), who describes <ll> and <y> as “somewhere between ‘ch’ and ‘zh’”, a sound “not pronounced in English”. Finally, Palmerlee et al. (2005: 489–496) call attention to “the trait of pronouncing the letters ll and y as ‘zh’ (as in ‘azure’)”, but then say that <y> is pronounced “as the ‘sh’ in ‘ship’ when used as a consonant” and consistently transcribe both <y> and <ll> with ‘sh’ (e.g. calle ‘street’ = “ka·she”, desayuno ‘breakfast’ = “de·sa·shoo·no”). These differing descriptions of Argentine palatals suggest that there is a high degree of variation in the phonetic realization of the palatal phoneme in Argentine Spanish. The nature of this variation is what is examined in the present study, which uses acoustic analysis to investigate the realization of palatals in a small sample of speakers from the Buenos Aires area. In the following sections, I review technical work that has been done on this topic, present new data on variation in palatal production and its sociolinguistic correlates, and conclude that a change from voiced to voiceless is already complete. 2. Background The sounds of Argentine Spanish have drawn the attention of many a Hispanicist over the last sixty years (e.g. Malmberg 1950; Honsa 1965; Fontanella de Weinberg 1983, 1985b, 1987, 1989; Donni de Mirande 2000), and the palatal consonants in particular have been the topic of several phonetic and phonological studies (e.g. Gandolfo 1964; Abadía de Quant 1988, 1996; Harris and Kaisse 1999; Baker * I am grateful to Daiana Chang, Ester Kim, and Daniel Kim for their help with recruiting participants, to all the porteños who were so kind as to read a little Mafalda, and to audiences at the Berkeley Phonetics and Phonology Forum and WSS 4 for their valuable comments and feedback. This work was supported in part by a Jacob K. Javits Fellowship and a National Science Foundation Graduate Research Fellowship. Any errors are mine alone. 55 UC Berkeley Phonology Lab Annual Report (2008) and Wiltshire 2003). With regard to Argentine palatals, the first linguistic feature that linguists point out is the early merger of what in European Spanish are (or used to be) two distinct phonemes, a palatal lateral /ʎ/ and a palatal glide /j/, a phenomenon called yeísmo (Fontanella de Weinberg 1985a, Guitarte 1992). The second feature that is commented upon is the occlusivization or spirantization that turns the merged sound’s non-sibilant realization into a sibilant realization, a phenomenon referred to as zheísmo. The result is that “Standard Buenos Aires has changed the phoneme /j/ into the voiced palatal fricative /ž/” (Honsa 1965: 277). Moreover, the “groove” (i.e. sibilant) fricative realization is noted by Vidal de Battini (1964) to have already become common in the northeastern and southern provinces of the Pampas, Mesopotamia, and Patagonia regions, though at that time non-sibilant palatals still predominated in the northwestern provinces of the Pampas, Argentine Northwest, and Cuyo regions. Indeed, later work by Lipski (1994: 170) claims that “[t]he groove fricative pronunciation of /y/ originated in Buenos Aires, from which it has spread steadily outward, representing the prestige standard which is associated with Argentine Spanish throughout the world”. The subject of this paper is a third phenomenon often referred to as sheísmo, the devoicing of a merged palatal phoneme that appears to have started in the colloquial speech of the lower working classes of Buenos Aires. Alonso (1961) indicates that fifty years ago voiced [ʒ] was the most widespread pronunciation in Buenos Aires, but that devoicing, while still relatively uncommon, was gaining currency. In fact, Honsa (1965: 278) estimates that the change from voiced /ʒ/ to voiceless /ʃ/ began approximately sixty years ago: [The colloquial Buenos Aires dialect] originated in the capital city at a time when large numbers of the peasant class fom [sic] the area were moving into Buenos Aires to join the ranks of the city workers, the supporters of Juan and Eva Perón. From all indications, a change in the dialect must have occurred between the years 1946 and 1949, at which time a phonemic mutation, the change of the voiced palatal fricative /ž/ to a voiceless /š/ took root and rapidly spread into the colloquial usage of all classes…The spread was greatly aided by the endlessly broadcast speeches of the Perón era and by the efforts of the higher classes to manifest to the pushing proletariat…that the higher classes were not anti-social. Though the initial innovation of [ʃ] may have appeared first in the speech of (middle-aged) workers, young people – young women in particular – appear to be largely responsible for its rapid spread. Resnick (1975) shows that in Buenos Aires, [ʒ] occurred in all social classes, while [ʃ] was colloquial and occurred only sometimes; however, the incidence of [ʃ] was especially high among younger speakers. Moreover, Canfield (1981) observes that while /ʎ/ and /j/ were generally leveled to [ʒ], [ʃ] occurred occasionally, especially in women’s speech. Finally, Honsa (1965: 279) notes that the regional supplanting of standard Buenos Aires dialect by colloquial Buenos Aires speech was “most noticeable in the usage of the young, no doubt powerfully influenced by the movies, records, radio, and television”. The first study to provide quantitative data on the coexistence of voiced and voiceless palatals in Argentina is Fontanella de Weinberg’s (1978) study of 60 speakers from one neighborhood in Buenos Aires. In this study [ʃ] was found to occur most often in females 15-30 years old, followed by females 31-50 years old and males 15-30 years old, and finally males 31-50 years old; [ʃ] was almost never found in speakers 51-70 years old.1 Thus, the conclusion is that devoicing began among young females, college-educated ones in particular, and then spread to other groups of speakers. Fontanella de Weinberg (1992) notes that voiced, devoiced, and completely voiceless realizations of the palatals coexist and, moreover, are socially conditioned, with the predominance of voiceless and devoiced variants being much greater in the youngest generation of speakers and amongst women.2 The findings of Fontanella de Weinberg are echoed by Lipski (1994: 170), who states that “[a]lthough the original sound was voiced [ʒ], most younger residents of Buenos Aires now pronounce 1 In later comments on this work, Fontanella de Weinberg (1992) states: “las mujeres menores de 30 años eran…el grupo que presentaba usos más ensordecidos…y les seguían las hablantes femeninas mayores de 30 años, mientras que los índices de ensordecimiento de los hombres eran mucho menores dentro de cada edad, aproximándose los varones de 15 a 30 años al uso de las mujeres de 31 a 70 años”. 2 Fontanella de Weinberg (1992) writes: “En la actualidad, coexisten realizaciones sonoras, ensordecidas y plenamente sordas…socialmente condicionadas, ya que entre los hablantes más jóvenes y entre las mujeres el predominio de las variantes sordas y ensordecidas es mucho mayor”. 56 UC Berkeley Phonology Lab Annual Report (2008) a voiceless [ʃ], and the devoicing is spreading throughout Argentina”, in no small part due to the prestige associated with the speech of the capital. Consequently, there are three types of systems described by Fontanella de Weinberg (1992): (i) that of older speakers, in which the palatals are voiced underlyingly and on the surface (/ʒ/ > [ʒ]); (ii) that of middle-aged speakers, in which the palatals are voiced underlyingly, but surface variably as either voiced or voiceless (/ʒ/ > [ʒ], [ʃ]); and (iii) that of younger speakers, in which the palatals are voiceless both underlyingly and on the surface (/ʃ/ > [ʃ]). In the present study, I examine whether these three systems can all be found in Buenos Aires today. What variants of the palatal phoneme are out there in Buenos Aires, and what is the sociolinguistic nature of the variation? The next two sections describe a production experiment conducted in June 2007 to get at answers to these questions.